EP0653237B1 - Process for reducing the concentration of nitrogen oxides in the exhaust gas of combustion engines or incinerators - Google Patents

Description

The invention relates to methods for reducing the nitrogen oxide concentration in the exhaust gas of an internal combustion engine or an incinerator according to the preamble of claim 1.

The nitrogen oxide and particle emissions (dust) of a diesel engine optimized for performance and consumption can only be reduced insignificantly by means of combustion technology. In order to be able to comply with the exhaust gas limit values prescribed by law in the future, aftertreatment of the diesel engine exhaust gases is therefore essential.

$${\text{2NO}}_{\text{2}}{\text{+ 4NH}}_{\text{3}}{\text{+ O}}_{\text{2}}{\text{→ 3N}}_{\text{2}}{\text{+ 6H}}_{\text{2}}\text{O}$$

ablaufen können. Zur Reduktion von 1 Mol NO_x im dieselmotorischen Abgas benötigt man etwa 0,9 bis 1,1 Mol NH₃. Wird weniger Ammoniak NH3 eingespritzt, arbeitet der Katalysator nicht mehr mit dem höchsten Wirkungsgrad. Eine Überdosierung ist ebenfalls zu vermeiden, da ansonsten unverbrauchtes Ammoniak NH3 in die Atmosphäre gelangt.A significant reduction in NO _x emissions from a diesel engine can be achieved by application of the so-called S elective- C atalytic- R eduction technique. In the SCR process, gaseous ammonia NH ₃ , ammonia in aqueous solution or urea as a reducing agent is injected into the exhaust system, so that the chemical reactions in particular on a catalyst $${\text{4NO + 4NH}}_{\text{3rd}}{\text{+ O}}_{\text{2nd}}{\text{→ 4N}}_{\text{2nd}}{\text{+ 6H}}_{\text{2nd}}\text{O}$$

$${\text{2NO}}_{\text{2nd}}{\text{+ 4NH}}_{\text{3rd}}{\text{+ O}}_{\text{2nd}}{\text{→ 3N}}_{\text{2nd}}{\text{+ 6H}}_{\text{2nd}}\text{O}$$

can expire. To reduce 1 mole of NO _x in the diesel engine exhaust, about 0.9 to 1.1 moles of NH _{3 are} required. If less ammonia NH3 is injected, the catalyst no longer works with the highest efficiency. Overdosing should also be avoided, as otherwise unused ammonia NH3 will get into the atmosphere.

The SCR process known from DE-A-36 10 364 can reduce the proportion of NOx in the exhaust gas from combustion plants by more than 80% and at the same time limit the NH ₃ emissions to below 5 ppm. The dosage of the reducing agent is monitored Computer which evaluates the output signal of an NH ₃ sensor arranged in the exhaust tract behind the NO _x converter and, if necessary, readjustes the supply of reducing agent by controlling a delivery unit. An NH ₃ sensor is an electrochemical cell which contains a cup-shaped body made of stabilized zirconium dioxide as an essential component. Two electrodes are applied to the solid electrolyte, the outer electrode exposed to the exhaust gas being made, for example, of TiO ₂ , Pt V ₂ O ₅ or V ₂ O ₅ and the inner electrode being exposed to a reference gas (air) made of platinum.

The process described in DE 36 06 535 for reducing the NO _x concentration uses the high temperatures of the exhaust gas to evaporate the reducing agent present in solid or liquid form and to split it into reactive components. After concentration, the reactive components are injected via a bundle of pipelines into a comparatively cool zone of the exhaust tract, where they react with the nitrogen oxides to nitrogen, water and carbon dioxide. A sensor working according to the chemiluminescence comparison method is used to measure the NO _x concentration. Its output signal is fed to a controller with setpoint specification, which controls the servomotor of a metering valve.

The aim of the invention is to provide a method with which the concentration of nitrogen oxides NO _x in the exhaust gas of an internal combustion engine or an incinerator can be significantly reduced. In particular, it should be ensured that the exhaust gas contains neither nitrogen monoxide NO nor excess ammonia NH ₃ . According to the invention, these objects are achieved by a method according to claim 1.

The advantage that can be achieved with the invention consists in particular in that the amount of NH ₃ required for a stoichiometric conversion of nitrogen monoxide NO to nitrogen N ₂ and water H ₂ O is obtained in a simple manner by measuring the electrical resistance of a sensor element exposed to the exhaust gas, preferably one Vanadate layer, can determine. An absolute measurement of the NO or NH ₃ concentration is not necessary. Since the aim of the control is a maximum resistance of the metal oxide layer used as the sensor element, any resistance drift that may be present does not cause any problems.

Fig. 1

eine mit einem SCR-Katalysator ausgestattete Abgasanlage eines Dieselmotors

Fig. 2 und 3

den schematischen Aufbau eines NO/NH₃-Detektors

Fig. 4

die Kammelektroden des NO/NH₃-Detektors

Fig. 5

Verfahrensschritte zur Herstellung der Kammelektroden

Fig. 6

eine auf den Kammelektroden abgeschiedene Al₂O₃-V₂O₅-Sandwich-Struktur

Fig. 7 bis 11

die Sensitivität einer AlVO₄-Dünnschicht auf Stickstoffmonoxid NO, Ammoniak NH₃ und andere Gase

Fig. 12

ein Ausführungsbeispiel eines Detektorgehäuses

Fig. 13

ein das Verfahren zur Regelung der eingespritzen Harnstoffmenge erläuternden Ablaufplan

Fig. 14

den elektrischen Widerstand einer Vanadatschicht in einem Stickstoffmonoxid NO und Ammoniak NH₃ enthaltenden Gasgemisch in Abhängigkeit von der Menge des dem Gasgemisch zugesetzten Ammoniaks NH₃

The dependent claims relate to advantageous developments and refinements of the invention explained below with reference to the drawings. Here shows:

Fig. 1

an exhaust system of a diesel engine equipped with an SCR catalytic converter

2 and 3

the schematic structure of a NO / NH ₃ detector

Fig. 4

the comb electrodes of the NO / NH ₃ detector

Fig. 5

Process steps for the production of the comb electrodes

Fig. 6

an Al ₂ O ₃ -V ₂ O ₅ sandwich structure deposited on the comb electrodes

7 to 11

the sensitivity of an AlVO ₄ thin film to nitrogen monoxide NO, ammonia NH ₃ and other gases

Fig. 12

an embodiment of a detector housing

Fig. 13

a flow chart explaining the process for controlling the amount of urea injected

Fig. 14

the electrical resistance of a vanadate layer in a gas mixture containing nitrogen monoxide NO and ammonia NH ₃ as a function of the amount of ammonia NH ₃ added to the gas mixture

The exhaust system of a diesel engine 1 shown schematically in FIG. 1 is said to largely break down the nitrogen oxides NOx formed during operation and to release the remaining residual gases to the atmosphere with as little noise as possible. It consists of an SCR catalytic converter 2 described in / 1 / and / 2 /, one or more switching dampers 3 and a pipe system 4 which connects the individual components to the exhaust gas outlet openings present in the cylinder head of the diesel engine 1. Upstream of the SCR catalytic converter 2 is a metering device 5, which is stored in a storage container 6 Injecting reducing agent into the exhaust tract. The metering device 5 contains in particular a membrane pump connected to an injection nozzle 7 or an injection valve with an upstream flow meter. A control unit 8 ensures that a certain amount of the ammonia-containing reducing agent can be supplied to the exhaust gas.

stöchiometrisch ablaufen. Erfindungsgemäß wird deshalb die NO- bzw. NH₃-Konzentration mit Hilfe eines im Abgasrohr 4 hinter dem SCR-Katalysator 2 angeordneten Detektors 9 gemessen und zur Steuerung der eingespritzten Harnstoffmenge herangezogen. Die für eine stöchiometrische Reaktion erforderliche Harnstoffmenge wird aufgrund der noch zu beschreibenden Eigenschaften des Detektors dann eingespritzt, wenn der elektrische Widerstand einer als NO- bzw. NH₃- sensitives Element verwendeten Vanadatschicht ein Maximum bzw. deren elektrische Leitfähigkeit ein Minimum durchläuft.An aqueous urea solution (CO (NH ₂ ) ₂ ) is particularly suitable as a reducing agent. This is decomposed with the addition of heat to carbon dioxide CO ₂ and ammonia NH ₃ , the ammonia NH ₃ adsorbing on the surface of the catalyst 2 and reacting with the nitrogen oxides NO and NO ₂ present in the exhaust gas to form the non-toxic substances nitrogen N ₂ and water H ₂ O. . The reaction must be carried out to ensure that the diesel exhaust gas entering the environment does not contain nitrogen monoxide NO or excess ammonia NH ₃ $${\text{4NH}}_{\text{3rd}}{\text{+ 4NO + O}}_{\text{2nd}}{\text{→ 4N}}_{\text{2nd}}{\text{+ 6H}}_{\text{2nd}}\text{O}$$

run stoichiometrically. According to the invention, the NO or NH ₃ concentration is therefore measured with the aid of a detector 9 arranged in the exhaust pipe 4 behind the SCR catalytic converter 2 and used to control the amount of urea injected. The amount of urea required for a stoichiometric reaction is then injected due to the properties of the detector to be described when the electrical resistance of a vanadate layer used as an NO or NH ₃ -sensitive element passes through a maximum or its electrical conductivity goes through a minimum.

The substrate 10 of the detector 9 shown in FIGS. 1 and 2 consists of an electrically insulating material such as glass, beryllium oxide BeO, aluminum oxide Al ₂ O ₃ or silicon (with Si ₃ N ₄ / SiO ₂ insulation). On the substrate 10, which is between 0.1 and 2 mm thick, there are two platinum electrodes 11, 11 'forming an interdigital structure, a vanadate layer 12 (AlVO ₄ or FeVO ₄ ) connecting these electrodes as NH ₃ or NO-sensitive element and a temperature sensor 13 arranged. The passivation layer made of silicon oxide, designated 14, shields the connecting leads 15, 15 'and 16, 16' respectively assigned to the two comb electrodes 11, 11 'and the temperature sensor 13 from the oxygen present in the exhaust gas.

In order to be able to set the desired operating temperature of up to 600 ° C. and to be able to keep it constant regardless of external influences, the detector 9 is actively heated with the aid of a resistance layer arranged on the rear side of the substrate 10. The resistance layer designated 17 in FIG. 2 consists, for example, of platinum (Pt), gold (Au) or an electrically conductive ceramic and has a meandering structure. Also shown is the approximately 10 to 100 nm thick metal layer 18 consisting of titanium (Ti), chromium (Cr), nickel (Ni) or tungsten (W), which improves the adhesion between the substrate 10 and the platinum electrodes 11, 11 ' .

The dimensions of the comb electrodes 11 and 11 'depend on the specific resistance of the sensor layer 12 applied over it in the desired temperature range. For example, the comb structure 11, 11 ′ can have thicknesses of 0.1 to 10 μm, widths of 1 to 1000 μm and electrode spacings of 1 to 100 μm. For a 1 µm thick AlVO ₄ layer 12, the following dimensions lead to well-measurable specific resistances in the temperature range between 500 and 600 ° C: electrode thickness D = 1.5 µm, length of the interdigital structure L = 1 mm, electrode spacing S = 50 µm.

FIG. 4 shows a scale representation of an interdigital structure in a top view. In this embodiment, a platinum resistance layer 19 is used as the temperature sensor. To produce the comb electrodes 11, 11 ', a 1.5 μm thick platinum layer 20 is first deposited on the heated corundum substrate 10 in a sputtering system (see FIGS. 5a, b). The structuring of the layer 20 takes place in a positive photo step, in which the photoresist 21 is applied at the location of the electrodes to be produced and exposed through a mask 22 (see FIG. 5c, d, e). The developed photoresist 21 protects the platinum layer 20 during the subsequent etching step (see FIG. 5f). After removing the photoresist 21 with acetone, the desired comb electrodes 11 and 11 '(see FIG. 5g) are obtained, on which the gas-sensitive vanadate layer 12 is subsequently deposited (see FIG. 5h).

The use of gold Au instead of platinum Pt as the electrode material has no influence on the gas sensitivity of the Al ₂ O ₃ / V ₂ O ₅ mixed oxide.

The extraordinary properties of the detector are based on the sputtering method to be used in the production of the gas-sensitive layer 12 and the subsequent tempering. The comb electrodes 11, 11 'can be coated, for example, in the Leybold Z490 sputtering system. Metallic vanadium (V) and aluminum (Al) serve as starting materials, which are reactively atomized from corresponding targets in a plasma consisting of 80% argon and 20% oxygen and are deposited on the heated substrate 10. The sandwich structure 23 shown in FIG. 6 is built up by alternately atomizing the two targets. It has a thickness of approximately 1 μm and consists of 60 to 80 approximately 10 to 15 nm thick V ₂ O ₅ or Al ₂ O ₃ layers, the Al ₂ O ₃ content being 50% to a maximum of 70% . The sputtering parameters are given in the following table. Residual gas pressure approx. 2 - 4 x 10 ^-6 mbar Sputtering gas pressure 4.2 x 10 ^-3 mbar Sputtering gas 20% O ₂ /80% Ar DC potential Al target: 155V V target: 225 V. Substrate temperature approx. 250 ° C

In order to produce a homogeneous mixed oxide, the sandwich structure 23 is annealed in air in a high-temperature furnace for about 5 to 15 hours. The furnace temperature has a decisive influence on the topography and the phase of the Al ₂ O ₃ / V ₂ O ₅ layers. Layers that have been tempered at temperatures T between 550 ° C ≦ T ≦ 610 ° C and consist of equal proportions of V ₂ O ₅ and Al ₂ O ₃ show an optimal sensitivity for ammonia NH ₃ and nitrogen monoxide NO. The aluminum vanadate AlVO ₄ , which is responsible for the high gas sensitivity, is created by tempering. The maximum working temperature of the vanadate layer is around 600 ° C. Aluminum vanadate AlVO ₄ has a triclinic unit cell with a = 0.6471 nm, b = 0.7742 nm, c = 0.9084 nm, α = 96.848 A, β = 105.825 A and χ = 101.399 A, whose volume V = 0, Is 4219 nm ³ .

Layers with an Al ₂ O ₃ content of more than 50% show a somewhat smaller measurement effect. However, they can still be used at higher temperatures of up to 680 ° C.

The following diagrams are intended to document the sensitivity or sensitivity of the AlVO ₄ thin layers produced by the described method to different gases. The size σ / σ ₀ (σ ₀ : conductivity of the sensitive layer in synthetic air (80% N ₂ /20% O ₂ )) is plotted as a function of the time t and the concentration of the respective gas.

Even the presence of the smallest amounts of nitrogen monoxide NO and ammonia NH ₃ in dry synthetic air leads to a significant increase in the conductivity of the aluminum vanadate AlVO ₄ (see FIGS. 7 and 8). The conductivity changes by about 75% if you add 10 ppm nitrogen monoxide NO to the air. The addition of 10 ppm ammonia NH ₃ increases the conductivity by more than a factor of 6.

As FIG. 9 shows, the specific resistance of the AlVO ₄ thin film increases in the presence of nitrogen dioxide NO ₂ . Since the aluminum vanadate shows a completely different behavior compared to nitrogen monoxide NO (reduction in the specific resistance, see FIG. 7), one can clearly differentiate between the two nitrogen oxides, provided only one of the two nitrogen oxides interacts with the sensitive element.

In addition to nitrogen monoxide NO and ammonia NH ₃ , the vanadate layer also responds to changes in the oxygen partial pressure and hydrogen H ₂ (see FIG. 9). The cross sensitivity to oxygen O ₂ and hydrogen H ₂ is, however, considerably smaller than the reaction to nitrogen monoxide NO and ammonia NH ₃ . For example, 500 ppm hydrogen H ₂ in air results in almost the same change in conductivity as the addition of 10 ppm nitrogen monoxide NO. The gases carbon monoxide CO (up to 1500 ppm), methane CH ₄ (up to 5000 ppm) and carbon dioxide CO ₂ (up to 1%) up to the concentrations given in brackets are not detectable. In a moist gas mixture (80 mbar H ₂ O), a clear decrease in the NH ₃ sensitivity is observed; however, it still remains twice as sensitive to nitrogen monoxide NO (see the right part of FIG. 10).

11 shows the sensitivity of the AlVO ₄ thin film in moist air (80 mbar H ₂ O) at 500 ° C. and a NO content of 10 ppm. Another gas in the specified concentration was added to the moist air within the time intervals marked by a horizontal line. The air therefore contained, for example, 1500 ppm carbon monoxide CO between the 60th and the 120th minute and an additional 10 ppm of nitrogen monoxide NO between the 80th and the 100th minute. As the measurement results show, the NO sensitivity of the AlVO ₄ layer is not influenced by the presence of carbon monoxide CO, methane CH ₄ and carbon dioxide CO ₂ . The addition of hydrogen H ₂ causes no masking of the NO sensitivity, but a clear cross sensitivity can be determined. A similar effect is observed with oxygen O ₂ when its concentration decreases from 20% to 2%.

The stainless steel housing shown in FIG. 12 is used to install the detector 9 in the wall of the exhaust pipe 4. The housing consists of two parts, the housing head 26 having a gas inlet opening 24 and a metal web 25 on the one with a bore 27 for receiving the detector 9 provided base body 29 is attached. Before welding the two parts 26 and 29, the detector 9 is glued in the bore 27 of the base body 29. After assembly, the sensitive element is located in an S-shaped curved flow channel, which connects the gas inlet opening 24 to the gas outlet opening 30. In the left part of FIG. 12, the ceramic plate 31 closing the bore 27 of the lower housing part 29 is also shown. It contains several channels through which the connecting wires 32 used for contacting the detector 9 are led to the outside.

With the aid of the NO / NH ₃ detector 9 described above, a simple and effective method for regulating the urea injection into the SCR catalytic converter 2 can be implemented. Since the detector 9 reacts to both nitrogen monoxide NO and ammonia NH ₃ by increasing the conductivity (see FIGS. 7 and 8), it is initially not possible to decide which of the two gases interacts with the sensitive layer 12. As can be seen from the flow chart shown in FIG. 13, the control unit 8 will cause the metering device 5 to first inject more urea into the exhaust gas. If this measure leads to an increase in the sensor resistance, the nitrogen monoxide NO may not yet have been completely converted to nitrogen N ₂ and water H ₂ O. The amount of urea injected is now increased until the sensor resistance reaches the maximum value indicated by an arrow in FIG. 14 and that the catalyst 2 leaving exhaust gas contains neither nitrogen monoxide NO nor excess ammonia NH ₃ .

However, if the multiple injection of urea leads to a lower sensor resistance, the excess NH ₃ in the exhaust gas must be reduced by reducing the amount of urea (see right part of the flow chart). In the diagram in FIG. 14, the maximum value of the sensor resistance defining the optimal amount of urea is therefore approached from the right.

The invention is of course not limited to the exemplary embodiments described. For example, it is possible to arrange a second detector based on sputtered Al ₂ O ₃ / V ₂ O ₅ layers in the exhaust tract in front of the injection nozzle 7. This detector is then preferably used to monitor the regulation described, by measuring the NO concentration and comparing it with the amount of urea injected in each case. The NH ₃ sensitivity of the detector does not have a disruptive effect here, since the engine exhaust gas in front of the injection nozzle 7 contains no ammonia NH ₃ .

Instead of urea, ammonia in aqueous solution or gaseous ammonia can also be used as the reducing agent, the reducing agent also being able to be injected directly into the SCR catalytic converter 2.

The method according to the invention can of course also be used in so-called DeNO _x systems for smoke gas denitrification (see, for example, / 3 /).

Literature:

• / 1 / Motortechnische Zeitschrift 49 (1988) 1, pp. 17 to 21
• / 2 / Motortechnische Zeitschrift 54 (1993) 6, pp. 310 to 315
• / 3 / Environment, 1986 No. 1, Technical report flue gas cleaning, FR 19 to FR 25

Claims (9)

1. Process for reducing the concentration of nitrogen oxides in the exhaust gas of combustion engines or incinerators, in which an ammonia-containing reducing agent is added to the exhaust gas and in which nitrogen oxides are converted to nitrogen and water in a catalyst unit (2) with exhaust gas flowing through it, characterized in

   - that a first sensor element (12) which reacts both to nitrogen monoxide and to ammonia with an increase in conductivity is disposed in the exhaust gas stream downstream of the catalyst unit (2),

   - that the resistance dependent on the nitrogen monoxide concentration and ammonia concentration or the electrical conductivity of the sensor element (12) is measured, and

   - that an amount of reducing agent is added to the exhaust gas which is such that the electrical resistance of the sensor element (12) is a maximum or the electrical conductivity is a minimum.
2. Process according to Claim 1, characterized in that gaseous ammonia, ammonia in aqueous solution or urea is added to the exhaust gas as reducing agent.
3. Process according to Claim 1 or 2, characterized in that the electrical resistance or the electrical conductivity is measured of a sensor element (12) composed of a metal oxide/vanadium oxide mixture.
4. Method according to Claim 3, characterized in that the electrical resistance or the electrical conductivity is measured of a sensor element (12) composed of an aluminium oxide/vanadium oxide mixture or an iron oxide/vanadium oxide mixture.
5. Process according to one of Claims 1 to 4, characterized in that the electrical resistance or the electrical conductivity is measured of a sensor element (12) composed of a vanadate MeVO₄, where Me denotes a trivalent metal.
6. Process according to Claim 5, characterized in that the electrical resistance or the electrical conductivity is measured of a sensor element (12) composed of aluminium vanadate or iron vanadate.
7. Process according to one of Claims 1 to 6, characterized in that a sensor element (12) is used which has a layer-type structure and is contacted by an electrode pair (11, 11').
8. Process according to one of Claims 1 to 7, characterized in that the sensor element (12) is actively heated and kept at a constant temperature.
9. Process according to one of Claims 1 to 8, characterized in that a second sensor element corresponding to the first sensor element (12) is disposed in the exhaust gas stream upstream of the catalyst unit (2) and is used to measure the nitrogen monoxide concentration.

Images